AI citations vs recs, AIO triggers, Gemini Siri, OKF & Google’s loyalty play.
Hi Marketers!
Welcome to the new issue of HighDegree*: Cutting Through the Noise in SEO & Digital Marketing
This issue is about the gap between getting mentioned and actually winning.
Lily Ray found that self-promotional “best of” listicles can backfire, handing AI Overviews a recommendation for your competitors instead of you. Ahrefs ran the numbers on when AI Overviews even show up in the first place. Apple brought a Gemini-powered Siri to WWDC with almost no detail on how websites get credit for what it pulls. Google quietly published a new content format built for AI agents, and it’s separately building a loyalty system that rewards publishers with an engaged, returning audience over one-off clicks.
Let’s jump in.
➞ Digital Marketing This Week
- Self-Promoting Listicles Are Backfiring in AI Search
- What Actually Triggers an AI Overview (and What Doesn’t)
- What Apple’s Gemini-Powered Siri Means for Your SEO Strategy
- Google’s New Open Knowledge Format Is Built for AI Agents
- Google Is Quietly Building an Audience Loyalty Machine
- Everything from Google Search This Week
- Everything from the Social Media and AI World
- Articles That Will Help You Refine Your Marketing Knowledge
➞ Subscribe
If you are not subscribed to HighDegree* yet, please subscribe to us by clicking the “Subscribe” button below. I will post every issue to HighDegree* substack version, so you can get it instantly in your inbox every time I hit publish!
➞ Self-Promoting Listicles Are Backfiring in AI Search
Lily Ray just published one of the more unsettling pieces of AI search research this year. She pulled responses and cited sources from 100 B2B “best [category]” queries in Google’s AI Overviews across three dates between April and June 2026, focused on self-promotional listicles where a company ranks itself number one in its own industry roundup.
The pattern she found: when a brand publishes its own “best X” listicle, Google still cites that listicle about as often as before, but recommends a different company instead 69% of the time. You write the article, Google reads it, and then names one of the competitors you mentioned as the actual answer.
Ray’s read is that this hits smaller and mid-sized brands hardest. Established, already-trusted brands can still get cited and recommended together. Everyone else is effectively handing AI a vote for whoever else they listed.
The fix isn’t to stop appearing in roundups, it’s to stop publishing your own. Let other sites rank you instead, and treat citations and recommendations as two separate metrics, because right now only one of them is converting.
Read the full report from lilyraynyc.substack.com ➞
If you’re not sure how often your own queries even trigger an AI Overview in the first place, Ahrefs just answered that with actual numbers.
➞ What Actually Triggers an AI Overview (and What Doesn’t)
Ahrefs analyzed 146 million SERPs across 86 different keyword traits to figure out exactly when Google decides to show an AI Overview, and the topline number is lower than most people assume: AI Overviews show up on 20.5% of all keywords, or roughly one in five searches, according to Ahrefs.
A few patterns stand out. Almost every AI Overview, 99.9% of them, appears on informational queries, and Ahrefs found they’re 1.9 times more common for non-branded searches than branded ones. Question-style queries trigger an AI Overview 57.9% of the time, and longer queries of seven words or more hit 46%.
Category matters a lot too. Science, health, and pets and animals queries see AI Overviews 35-44% of the time. Shopping sits at just 3.2%, real estate at 5.8%, and local search at 7.9%. If most of your traffic comes from local or transactional terms, AI Overviews probably aren’t your biggest threat right now. If you’re writing informational, question-based content, they almost certainly are.
The practical move: pull your own keyword list and check which pages fall into the high-risk categories, then prioritize those for AI Overview tracking instead of treating every page as equally exposed.
Read the full report from ahrefs.com ➞
Google’s AI Overviews aren’t the only place this fight is happening. Apple just brought a very similar fight to the iPhone.
➞ What Apple’s Gemini-Powered Siri Means for Your SEO Strategy
Apple introduced its Gemini-powered Siri at WWDC this year, and according to Search Engine Journal’s Matt G. Southern, two details matter more than anything else in the keynote. Siri can now pull current information from the web and generate answers on practically any topic, and it’s built directly into Spotlight on iPad and Mac, where people already type questions instead of speaking them.
What Apple hasn’t said is what happens to the websites Siri pulls from. The most specific thing Southern could find was an updated Applebot support page stating that web answers may include links to sources, with no detail on when those links appear, how often, or how anyone outside Apple would measure it. A page could show up in Siri’s answers every day or never, and from the outside you’d have no way to tell the difference.
That’s the real takeaway here. The rollout happens in stages, so there’s time to prepare, but right now there’s no analytics dashboard, no Search Console equivalent, and no public signal. Make sure your content is structured and citable everywhere else, because Siri is pulling from the same pool of trusted, well-structured sources that already win in Google’s AI Overviews.
Read the full report from searchenginejournal.com ➞
While Apple stays quiet about source attribution, Google just published a new spec that’s much more transparent about how it wants agents to read your content.
➞ Google’s New Open Knowledge Format Is Built for AI Agents
Google quietly published documentation for something it calls the Open Knowledge Format (OKF), and Marie Haynes thinks it’s a bigger deal than it sounds. An OKF bundle is just a directory of Markdown files, but instead of turning each webpage into one file, you break your content down into individual concepts. A single page might produce several separate knowledge files.
Each file opens with YAML frontmatter, a short block of metadata listing the content type, title, and tags, so an agent can tell what’s inside without reading the whole document first. The body underneath can be anything: headings, lists, tables, code blocks, actual process documentation.
Haynes connects this to the LLM Wiki pattern, an idea credited to AI researcher Andrej Karpathy, where an agent builds and maintains a standing wiki of knowledge instead of repeatedly retrieving and searching through raw documents.
If this catches on, the skill that matters shifts again: from writing pages that rank to organizing a business’s knowledge into something an agent can use directly. Haynes suspects this isn’t quite SEO, GEO, or even agent optimization, and that it might need its own name. Businesses that build OKF bundles early will have a head start figuring out what that name should be.
Read the full report from mariehaynes.com ➞
OKF is about reshaping content for agents. Google’s other recent moves are reshaping its relationship with human readers, and that pattern is just as worth watching.
➞ Google Is Quietly Building an Audience Loyalty Machine
Barry Adams spotted a pattern running through several recent Google features that, on their own, look like unrelated updates. Put together, he argues, they add up to Google building a loyalty system for publishers.
Preferred Sources, rolled out globally in April 2026, lets users pick specific publishers they want to see more of, and since May those picks show up inside AI Overviews and AI Mode, not just Top Stories. Search Profiles give publishers and creators with over 100,000 followers a dedicated page that users can follow for more visibility in Discover. Subscription Linking connects a reader’s existing subscription to their Google account, so subscribed content gets surfaced more prominently across search, Discover, and now AI surfaces too.
None of these bring back the cheap, high-bounce traffic that AI Overviews have eaten into. Adams’s point is that they were never meant to. Google has spent two decades replacing low-value content with direct answers, and generative AI just finished the job. What’s left is a system that rewards publishers who already have an engaged, loyal audience instead of those chasing one-off clicks.
The strategic shift: stop measuring success by traffic, and start measuring it by how many readers actively choose to come back.
Read the full report from seoforgooglenews.com ➞
➞ From Google
Everything from Google search this week –
→ Google: LLMS.txt Files Won’t Help or Hurt Your Search Rankings (seroundtable.com)
→ Google: No Practical SEO Difference by Going with Folder for US Site (seroundtable.com)
→ Headline Formats and Google Discover: What 3.4 Million Articles Reveal (searchengineland.com)
→ 97% of LLMS.txt Files Got No Requests, Ahrefs Data Shows (searchenginejournal.com)
→ Google Ads to Limit Ad Impressions from Unqualified Advertisers (seroundtable.com)
→ Google Ads Bidding Target Optimization Changing, Promotion Mode Beta, Smart Bidding Exploration Expands (seroundtable.com)
→ Google Shopping Results with Good Price Label (seroundtable.com)
→ Google Business Profile Owner/Manager Access Invites Not Working for Some (seroundtable.com)
→ Google Local Finder Interface Drops Pagination (seroundtable.com)
→ Google Search Rolls Out Information Agents in AI Mode for a Fee (seroundtable.com)
→ Google Can Be Directly Liable for False AI Overview Claims: German Court (searchengineland.com)
→ Gemini Can Help with Your Google Business Profiles (seroundtable.com)
→ WhatsApp Numbers Added to Google Business Profiles in Bulk (seroundtable.com)
→ Google Partners with Walmart Connect, Updates Real Estate Listings (socialmediatoday.com)
→ Enhanced Local Services Ads for Home Listings Bring Homebuyers and Local Agents Together (blog.google)
→ Google Ads Sitelinks with Images (seroundtable.com)
→ Ginny Marvin Clarifies AI Max, AI Search Ads, and What Advertisers Should Prioritize after GML (searchengineland.com)
→ Google Merchant Center Temporary Delay on Product Submissions (seroundtable.com)
→ Google Tests Dotted Underlined and Shaded URLs in Search Snippet (seroundtable.com)
→ Google Search Was Unaffected by Gemini Outage (seroundtable.com)
→ Google Will Save Your Lens Photos, Search Live Recordings, and Translate Audio for AI Training (theverge.com)
➞ AI + Social
Find out what’s happening in the social media and artificial intelligence world –
→ ChatGPT’s Market Share Slips below 50% for the First Time (techcrunch.com) – OpenAI’s chatbot lost majority share of the AI assistant market as competitors gained ground.
→ Anthropic Forced to Shut Down Fable 5 by U.S. Government Order (searchenginejournal.com) – A government directive pulled Anthropic’s Fable 5 model despite the company’s objections.
→ Zhipu Surges 33% as Wall Street Raises Bets on China AI after Anthropic Curbs (cnbc.com) – Chinese AI stocks rallied after U.S. restrictions on Anthropic shifted investor attention toward domestic alternatives.
→ Claude Doesn’t Re-Rank Search Results, It Transposes Brave’s Top 10 Directly (linkedin.com) – A marketer’s analysis suggests ranking well on Brave search may directly determine whether you show up in Claude’s answers.
→ Bing Webmaster Tools Updates AI Reporting with Intents, Topics, Citation Share, and Compare (searchengineland.com) – Microsoft added new reporting views that break down AI search performance by intent and citation share.
→ Microsoft Weighs DeepSeek for Copilot Cowork (axios.com) – Microsoft is reportedly considering DeepSeek’s model to power parts of its Copilot Cowork product.
→ New Microsoft Advertising Product Explorer (seroundtable.com) – Microsoft rolled out a new tool for browsing and managing product feed data in its ad platform.
→ Meta’s New “AI Mode” on Facebook Pulls from Public Info across Its Platforms (techcrunch.com) – Facebook’s new AI assistant feature draws on public data from across Meta’s apps to answer questions.
→ Meta Announces World Cup Features across Its Apps (socialmediatoday.com) – Meta is rolling out World Cup-themed tools and content formats across Facebook, Instagram, and WhatsApp.
→ Meta’s Threads Reaches 500 Million Users Three Years after Launch (bloomberg.com) – Threads hit half a billion users, a milestone that puts real scale behind Meta’s text-based platform.
→ Edits Is Getting a Desktop Version and an AI Production Assistant (socialmediatoday.com) – Meta’s video editing app for creators is expanding to desktop with built-in AI editing help.
→ TikTok Launches Digital Trading Cards for the 2026 FIFA World Cup (socialmediatoday.com) – TikTok is testing a collectible card feature tied to World Cup content and creators.
→ Reddit Campaigns Drive Performance on and off the Platform (socialmediatoday.com) – New data suggests Reddit ad campaigns are influencing purchase behavior beyond the platform itself.
→ Pinterest Adds AI Updates for SMBs (socialmediatoday.com) – Pinterest rolled out new AI tools aimed at helping small businesses create and manage content.
→ YouTube Announces New Live Concert Series (socialmediatoday.com) – YouTube is launching a recurring live concert format to draw more real-time viewership.
→ Anthropic Walks Back Policy That Could Have “Sabotaged” AI Researchers Using Claude (wired.com) – Anthropic reversed a policy change after backlash from researchers who said it would have disrupted their work.
→ Cybersecurity Researchers Aren’t Happy about the Guardrails on Anthropic’s Fable (techcrunch.com) – Security professionals say Fable’s added safety restrictions are getting in the way of legitimate research.
→ Claude Fable Won’t Answer Basic Biology Questions (theverge.com) – Anthropic’s Fable model is declining to answer routine biology questions due to its new safety guardrails.
→ Anthropic’s Fable 5 Can Make Weirdly Fun Video Games with the Click of a Button (techcrunch.com) – Despite the controversy around its guardrails, Fable 5 is drawing attention for generating playable games from simple prompts.
→ Claude Is Now Driving 386% More Traffic to Websites, but Its Role in Workflow Automation Could Matter More (seranking.com) – New research shows referral traffic from Claude has jumped sharply, though the bigger shift may be in how Claude automates marketing workflows.
→ OpenAI Launches Product Feed Ads in Ads Manager Beta (searchengineland.com) – OpenAI is testing a new ad format that lets advertisers submit product feeds directly into its ad platform.
→ Amazon Turns Alexa into a Shopping Agent and an Advertising Platform (searchengineland.com) – Amazon is repositioning Alexa to handle purchases directly and to carry advertising for brands.
→ DoorDash’s New AI Chatbot Lets You Order with Prompts and Photos (techcrunch.com) – DoorDash added an AI ordering assistant that can interpret text prompts and images to place food orders.
→ Meta Will Use Off-Site Activity to Customize Feed and AI Responses (socialmediatoday.com) – Meta confirmed it will pull in browsing activity from outside its apps to personalize feeds and AI answers.
→ Meta Publishes 2026 Holiday Planning Guide (socialmediatoday.com) – Meta released its annual guide for brands planning holiday season ad campaigns.
→ Instagram Extends Your Algorithm to the Main Feed (socialmediatoday.com) – Instagram’s user-controlled recommendation tool is now available on the primary feed, not just Explore.
→ LinkedIn Launches Its Own Creator Marketplace (socialmediatoday.com) – LinkedIn introduced a marketplace connecting brands with creators directly on the platform.
→ LinkedIn Outlines Effective B2B Marketing Strategies (socialmediatoday.com) – LinkedIn shared new guidance on what’s working for B2B marketers on the platform right now.
→ LinkedIn Launches Its Eighth Puzzle Game (socialmediatoday.com) – LinkedIn added another daily puzzle game as part of its push to boost everyday engagement.
→ Scroll to Shop: What Drives Real FMCG Growth on TikTok (ads.tiktok.com) – New research from TikTok and Circana breaks down which content patterns actually drive consumer goods sales.
→ Pinterest Adds Amazon Storefront Linking (socialmediatoday.com) – Pinterest now lets sellers link directly to their Amazon storefronts from pins.
→ YouTube Expands In-App Messaging (socialmediatoday.com) – YouTube is rolling its direct messaging feature out to more users.
➞ Worth Reading
These are the articles that will help you refine your marketing knowledge –
→ Should I Use Markdown for My Site? Google Search Central (youtube.com)
→ We Analyzed 137K Sites: 97% of LLMs.txt Files Never Get Read (ahrefs.com)
→ In 2026, Less than One Third of Google Searches Still Send a Click (sparktoro.com)
→ Why Your Product Feed Is an SEO Asset (and Who Should Own It) (searchenginejournal.com)
→ The New Rules of Local Search: Reviews, Community Signals, and Brand Prominence (advancedwebranking.com)
→ How to Build a Representative AI Search Prompt Library for Better AI Visibility Measurement (aleydasolis.com)
→ Same Users, Same Jobs, Different Doors: How Organic and AI Search Cover the Same Job Universe (brightedge.com)
→ SEO Panel Agrees: Brand Is the New Backlink for AI SEO (searchenginejournal.com)
➞ What This All Means for Your Strategy
Every story in this issue points at the same shift. Being mentioned by Google’s AI isn’t the same as being chosen by it, whether that’s an AI Overview recommending a competitor from your own listicle, Siri quietly deciding which sites get a link, or Google rewarding publishers with an existing loyal readership over those still chasing clicks.
The brands that win going forward are the ones that already have trust built up somewhere: a name other sites recommend without being asked, a clear and well-structured page Apple’s web answers can pull from confidently, a knowledge base organized for the agents now reading it, and a reader base that picks you on purpose. None of that happens overnight, but all four are buildable starting now.
Until next week,
Nishat from HighDegree* Marketing
Thanks for reading HighDegree* Marketing! This post is public, so feel free to share it.
P.S. Have a question about implementing these strategies? Hit reply – we read every email and often feature reader questions in future issues.
➞ Who is Nishat Shahriyar?
I’ve been a Digital Marketing Strategist since 2007. Today I’m the Product Marketing Lead at Fluent Forms (a WordPress form and lead-generation plugin used by 700K people), FluentPlayer (a WordPress Video Player Plugin) and Ninja Tables (a WordPress table builder plugin used by 80K people). Before that, I led marketing at Fluent Support (a WordPress helpdesk plugin used by 10K people).
Connect with me on LinkedIn or follow me on X/Twitter at @rednishat for SEO & Marketing updates.